quant-ph digest — 2026-05-10
Scored against Yuan's research programme (Y1–Y6):
- Y1 — arXiv:2502.09704 — iterative warm-started QAOA
- Y2 — arXiv:2304.06915 — quasi-binary portfolio QAOA
- Y3 — arXiv:2410.16265 — QAOA DGMVP portfolio (QST 2026)
- Y4 — arXiv:2603.14744 — Grover + ADMM cardinality-constrained BO
- Y5 — arXiv:2510.08292 — GW speed-ups via Gibbs states + Pauli sparsity
- Y6 — arXiv:2510.11213 — PBR test on IBM Heron2
Source
arXiv listing: https://arxiv.org/list/quant-ph/new (55 new + 18 cross = 73 entries; announce cycle Friday 8 May 2026)
Coverage: all 73 entries scored. 7 relevant (score ≥ 1); 66 SKIP (score 0, omitted).
Scoring rubric
0–10 on method/scope/conclusion overlap — max wins. HIGH 8–10 · MED 5–7 · LOW 1–4 · SKIP 0.
Highly relevant (score 8–10) — 0 papers
No HIGH-scoring papers in today's announce cycle. No deep-analysis pass run.
Moderately relevant (score 5–7) — 2 papers
Quantum-enhanced Large Language Models on Quantum Hardware via Cayley Unitary Adapters
- Authors: Borja Aizpurua, Sukhbinder Singh, Augustine Kshetrimayum, Saeed S. Jahromi, Roman Orus
- arXiv: 2605.05914
- Category: new submission — Quantum Physics (quant-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
- Score: 5/10 (MED)
- Overlaps with: Y6 — scope overlap (same 156-qubit IBM Quantum System Two superconducting hardware family that hosts Heron r2)
- Why it matters: Y6 ran the PBR test on IBM's 156-qubit Heron2 in System Two; this paper runs an LLM-fine-tuning workload on the same hardware platform and reports a sharp noise–expressivity phase transition. The noise characterisation methodology and any device-specific calibration choices may be directly relevant for interpreting Heron-class qubit-pair performance in future Yuan experiments.
Large language models (LLMs) have transformed artificial intelligence, yet classical architectures impose a fundamental constraint: every trainable parameter demands classical memory that scales unfavourably with model size. Quantum computing offers a qualitatively different pathway, but practical demonstrations on real hardware have remained elusive for models of practical relevance. Here we show that Cayley-parameterised unitary adapters -- quantum circuit blocks inserted into the frozen projection layers of pre-trained LLMs and executed on a 156-qubit IBM Quantum System Two superconducting processor -- improve the perplexity of Llama 3.1 8B, an 8-billion-parameter model in widespread use, by 1.4% with only 6,000 additional parameters and end-to-end inference validated on real Quantum Processing Unit (QPU).
Architecture Shape Governs QNN Trainability: Jacobian Null Space Growth and Parameter Efficiency
- Authors: Michael Poppel, David Bucher, Maximilian Zorn, Markus Baumann, Sebastian Wölckert, Claudia Linnhoff-Popien, Philipp Altmann, Jonas Stein
- arXiv: 2605.05942
- Category: new submission — Quantum Physics (quant-ph); Machine Learning (cs.LG)
- Score: 5/10 (MED)
- Overlaps with: Y1, Y3 — method overlap (variational ansatz parameter trainability, structural rank deficiency / barren-plateau-adjacent phenomenon directly affects QAOA layerwise optimisation)
- Why it matters: Y3 reported that layerwise QAOA optimisation was the most robust strategy on noiseless DGMVP instances; Y1 likewise depends on warm-started parameter optimisation behaving well. This paper proves a hard rank bound (rank(J) ≤ 2L+1) on serial single-qubit ansätze and quantifies "structural gradient starvation" — both directly explain when adding more variational parameters helps vs. wastes optimiser budget. Worth checking whether the same Jacobian-rank diagnostic applies to the QAOA mixer–cost layered structure used in Y3.
Variational quantum circuits with angle encoding implement truncated Fourier series, and architectures arranging $N$ qubits with $L$ encoding layers each -- sharing encoding budget $E = NL$ -- generate identical frequency spectra, identical frequency redundancy, and require the same minimum parameter count for coefficient control. Despite this equivalence, trainability varies substantially with architecture shape $(N,L)$ at fixed $E$. We identify structural rank deficiency of the coefficient matching Jacobian $J$ as the mechanism responsible. For serial single-qubit architectures, we prove $\mathrm{rank}(J) \leq 2L+1$ regardless of parameter count $P$, with $\dim(\ker J) \geq P-(2L+1)$ growing without bound — a phenomenon we term structural gradient starvation: a growing fraction of parameters become structurally decoupled from the loss as $P$ increases at fixed $L$.
Tangential (score 1–4) — 5 papers
- 2605.05477 · score 3/10 · Operationally Admissible Post-Quantum Correlations from a Standard Quantum Walk — foundations adjacency to Y6 (PBR ontic/epistemic vs. CHSH/Tsirelson axis); shows a quantum walk with extended coin preparation can exceed the Tsirelson bound in coin-position correlations.
- 2605.05479 · score 3/10 · Quantum Simulation of the Real-time Dynamics in the multi-flavor Gross-Neveu Model at the utility scale using Superconducting Quantum Computers — IBM superconducting NISQ hardware (Y6 scope); LDOA method for diagonal unitary synthesis is generic enough to inform circuit design beyond field theory.
- 2605.06122 · score 3/10 · Variationally Compressing Quantum Circuits to Approximate Nonadiabatic Molecular Quantum Dynamics — variational circuit compression / hybrid quantum-classical optimisation method-adjacent to QAOA but for chemistry, not combinatorial optimisation.
- 2605.06167 · score 2/10 · Matrix encoding method in variational algorithm of calculating eigenvalues and generalized eigenvalues — variational/gradient method for matrix eigenvalues; tangential to QAOA-style hybrid optimisation.
- 2605.06224 · score 2/10 · Modular wedge localization, Majorana fields and the Tsirelson limit of the Bell-CHSH inequality — relativistic-QFT foundations adjacency to Y6 (Bell-CHSH near Tsirelson, distinct from the PBR ontic/epistemic axis).
Summary table
| Score | arXiv ID | Short title | Overlaps | arXiv |
|---|---|---|---|---|
| 5 | 2605.05914 | LLMs via Cayley unitary adapters on IBM 156-qubit System Two | Y6 (scope: hardware) | link |
| 5 | 2605.05942 | Architecture shape governs QNN trainability (Jacobian null-space) | Y1, Y3 (method) | link |
| 3 | 2605.05477 | Post-quantum correlations from a standard quantum walk | Y6 (foundations) | link |
| 3 | 2605.05479 | Gross-Neveu real-time dynamics on IBM superconducting (utility scale) | Y6 (scope: NISQ) | link |
| 3 | 2605.06122 | Variationally compressing Trotter circuits for chemistry | Y3 (method) | link |
| 2 | 2605.06167 | Variational algorithm for matrix eigenvalues | Y3 (method) | link |
| 2 | 2605.06224 | Modular wedge, Majorana, Tsirelson limit of Bell-CHSH | Y6 (foundations) | link |